Einstein AI is a critical lever within Salesforce B2C Commerce for driving sales through highly personalised experiences. From product recommendations to dynamic sorting and intelligent search, these capabilities are built into the platform — but only deliver value when properly configured. This article covers the five core Einstein features and what you need to know to get them working.
Commerce Insights: Understanding Shopper Behaviour
Einstein collects anonymous behavioural data through clickstreams — capturing user interactions such as product views, recommendation clicks, searches, and cart activity. This data feeds the personalisation engine.
Important: Activity tracking is disabled by default. To enable it, navigate to Merchant Tools > Site Preferences > Privacy Settings. Without this, Einstein has no behavioural data to work with and personalisation will be minimal.
Einstein Search Dictionaries
Search dictionaries improve how the platform understands shopper intent. They are enabled by default but initially use only site-specific data.
To access community data — which draws from aggregated, anonymised data across other B2C Commerce merchants — you must accept the Einstein Data Privacy Agreement at Administration > Global Preferences > Einstein Search Dictionaries Opt-In. This significantly enhances search accuracy.
Einstein Search Recommendations
This feature personalises the search-as-you-type experience by surfacing suggestions based on each individual shopper’s preferences and browsing history. Rather than showing the same suggestions to everyone, it tailors the typeahead to each user.
Implementation requires development effort to customise the search interface, but there are no specific Business Manager configurations required — the feature works from the behavioural data once the development work is in place.
Einstein Predictive Sort
Standard sorting attributes — price, new arrivals, best sellers — treat all shoppers the same. Predictive Sort goes further, intelligently ranking products based on individual shopper behaviour analysis.
To configure:
- Select a sorting rule in Business Manager
- Add the Predictive Sort attribute to the rule
- Set text relevancy parameters
- Apply the changes
The result is a product listing page that presents different customers with products in a different order — the order most likely to drive a purchase for that specific individual.
Einstein Product Recommendations
Product recommendations use real-time insights from clickstreams and purchase history rather than static rules or manual curation. They integrate into storefronts through content slots configured within the Einstein Configurator in Business Manager.
Done well, product recommendations drive meaningful improvements in conversion rate and average order value. The key is ensuring the clickstream data is flowing (see Commerce Insights above) and that recommendation slots are placed where they have genuine influence — product detail pages, cart pages, and category pages.
Putting It Together
Einstein AI in SFCC is not a switch you flip — it is a system that requires proper data collection, configuration, and ongoing attention. The five features above build on each other: accurate clickstream data improves search recommendations, which improves predictive sort, which improves the quality of product recommendations. Start with enabling activity tracking, and build from there.
